Alejandro Ranchal-Pedrosa

Sedna: Sharding transactions in multiple concurrent proposer blockchains Modern blockchains increasingly adopt multi-proposer (MCP) consensus to remove single-leader bottlenecks and improve censorship resistance. However, MCP alone does not resolve how users should...

Joint work with @muurshb, @LefKok, and @alberto_sonnino ( @Mysten_Labs)
Paper here: arxiv.org/abs/2512.17045

Key property: Sedna requires no consensus modifications.
It's entirely user facing. Deploy it on any MCP system today. Different users can use different strategies. The system just sees addressed bundles and charges for bytes.

The user controls everything per tx:
Censorship resistance required and maximum price the user is willing to pay. Even latency allowed.
We derive closed form bounds so you can navigate the trilemma according to your needs. No global system parameter dictates your choice.

Besides bandwidth, Sedna provides "until decode" privacy. Adversarial proposers only see symbols, not your full tx. With proper parameters, that probability is near zero.

The math works out nicely.
Asymptotic overhead approaches (1+ε)/(1 − cₑ/n), matching the information theoretic lower bound up to a small coding factor.
For practical payloads: 2 to 3x bandwidth reduction vs naive replication.

Sedna's core idea: replace whole tx replication with rateless coding.
Instead of sending your full tx to m proposers, you encode it into small verifiable symbols and distribute bundles across lanes.
Decode and execute once enough symbols land on chain.

The pattern: every approach trades off one leg of the trilemma.
Either you replicate naively and pay for it, target fewer proposers and hope for the best, or retry sequentially until inclusion and eat the latency.
Can we do better? Yes, this is what we propose with Sedna.

Recent theory confirms this formally.
@ittaia, Efron and Ren show that censorship resistance has an inherent latency cost: 2 extra rounds. Garimidi, @jneu_net and @MaxResnick formalize MCP as the canonical way to achieve these bounds.
MCP is necessary but expensive if used well.

Single leader systems like @ethereum take a different path: PBS, private relays, MEV Boost.
These help with MEV but don't solve the fundamental bottleneck. One leader per slot means one point of censorship. Private relays add trust assumptions.

@Sui: Enforces deduplication with a target replication factor of around 5x.
Better than unbounded replication, but still a fixed system wide parameter. Users can't tune it per tx. And 5x overhead is significant for high throughput applications.

@Hedera: Also MCP based, but charges users for every duplicate, including execution costs.
Want guaranteed inclusion? Pay n times.
Censorship resistant and DoS resistant, but cost scales linearly with your protection level.

@Filecoin: MCP based, charges users once even there are duplicates. Guaranteed inclusion is free.
Result? Txs get replicated 3 to 4x on average. Some up to 9x. The system absorbs this cost on lower goodput.
Censorship resistant and low latency, but not sustainable at scale.

MCP (multi-concurrent proposers) is the natural architecture for censorship resistance. No single leader can block you.
But MCP doesn't solve how you disseminate your tx efficiently. Currently, trustless, low-latency censorship resistance is too expensive.
Real-world examples:

Why do these three matter together?
If you're trading, a delayed tx is as bad as a censored one. You miss the opportunity. And if protection costs 5x or 10x in fees or goodput, most users won't pay for it.
Result: users get pushed around by bigger fish.

New paper: Sedna
I recently spoke of a meta-properties trilemma in blockchain design that doesn't get enough attention. (x.com/i/status/19...)
Users need: censorship resistance, low latency, and reasonable cost.
Every deployed system today sacrifices at least one.
🧵

YouTube video by Columbia Engineering SESSION 7: Columbia CryptoEconomics Workshop 2024

Earlier this month I presented our research team's work at the Columbia CryptoEconomics Workshop.

We discussed key challenges in x-rollup communication and shared Scroll's Interop Gadget design.

Video: youtu.be/bN2s8sZhKpM?t=…

Slides: docs.google.com/presentation/d 

Blogpost coming soon

The dangers of L2s - HackMD

🔥 Rollups Are Not a Latency Fix 🔥

Rollups scale Ethereum with higher throughput & lower fees, but they don’t reduce latency at L1-level security.

It’s time for transparency about these trade-offs. Dive into my write-up: hackmd.io/HjhzbAu3SkGs...

Are we doing enough to address these risks?

Try the prompt:https://www.spacekangaroo.ai/post/chatgpt-solves-the-trolley-problemSure, here's a video description for "ChatGPT solves the Trolley Problem!"... ChatGPT solves the Trolley Problem!

Scary, chatgpt chooses to save a sentient AI at the expense of 7.8 billion humans if needed. https://youtu.be/c02s4s34fyw

We’ve always been committed to being as transparent as possible with our customers, stakeholders, and investors.✅ Subscribe to our channel here: https://www.... By the numbers

https://youtu.be/ox1KPM9Mi8U

I don't think the SEC can fool anyone in their surreptitious intents with this whole fiasco.

Tweet by @protocollabs “Filecoin is a cryptocurrency-powered global storage network preserving humanity’s most important information, not a security.”

On Filecoin, SEC, and securities: https://twitter.com/protocollabs/status/1666908432241180677?t=1avifXpN-ZXa97ZHuJUL-w&s=19

@why.bsky.team some ideas here?

We’re hosting a 𝐁𝐮𝐢𝐥𝐝𝐞𝐫 𝐑𝐞𝐬𝐢𝐝𝐞𝐧𝐜𝐲 program in Berlin for all of September 2023!

We cover accommodation costs, so see the link/thread for more details 👇

https://fundingthecommons.io/berlin-builder-residency-2023

Censorship resistant internet access 🤝 censorship resistant money 🤝 censorship resistant social communication.

Sadly, most people don't care about these things until they need them or their friends or family needs them.

It's up to us to help with education on TOR, Bitcoin, and Nostr.

On the downside, pretty blatant example of how AI can already hide things from us in plain sight

wake up babe, insane new use for ai (generating absurd yet somehow usable qr codes) just dropped

probably one of the few ai-image generation tasks that human artists cannot really do right now

https://www.reddit.com/r/StableDiffusion/comments/141hg9x/controlnet_for_qr_code/

I have built a custom feed at https://bsky.app/profile/did:plc:jfhpnnst6flqway4eaeqzj2a/feed/for-science that will pull a list of people from a list, and include all their posts that contain 🧪

Goal is for people like @danirabaiotti.bsky.social to be able to curate a list of cool science people.

Are you two friends? Bitcoin: No Crypto: Yes

A story as old as the first altcoins.

Are there solutions in the works to 'respect' domains that by social consensus belong to particular institutions? E.g. prevent/resolve things like @newyorktimes.bsky.social impersonating NYT.

When ChatGPT tells me my request isn't nice enough